One of the most significant recent developments in search is faceted search. If you’ve used a modern site search or enterprise search system, you’ve likely stumbled into faceted search. After reading this article, you’ll be sure to recognize it when you see it. Moreover, if you work at a company that needs to make its information findable, you’ll likely include faceted search in the requirements for any search applications you or your company develops.

Faceted search is a technique for accessing a collection of information represented using a faceted classification scheme.

In a faceted classification scheme, each object or document can be classified based on multiple facets that represent independent ways of describing it. For example, the facets in a product catalog include category, brand, price, etc. The facets in a library catalog include subject, author, time period, etc. Faceted classification represents the fact that, as David Weinberger says, everything is miscellaneous, and it is artificial (if not impossible!) to arrange a large, diverse collection into single, pre-determined, hierarchy or taxonomy.

The screenshots below from the Home Depot and the State University Libraries of Florida web sites that provide faceted search interfaces. In these examples, the facets are displayed on the left-hand side, but every site takes its own design approach. To get an idea for the possibilities, check out Peter Morville’s collection of faceted navigation design patterns.

<Please allow time to load – the images are important to understanding!>

  

As a practical matter, the facets for a collection are typically derived either from pre-existing fields in a database (e.g., the columns in a relational database or data warehouse) or by applying information extraction techniques to unstructured content, (e.g., detecting names of people and places in free text). 

Recent years have seen an accelerated interest in faceted search, both by library and information scientists in the academic community and by practitioners in industry.

One of the leading academic researchers in this area is Marti Hearst, who leads up the Flamenco project at the University of California, Berkeley. She not only has led the development of prototype systems, but has also performed numerous usability studies to drive design recommendations for faceted search implementations.

But the biggest success of faceted search has been its adoption by the mass market, particularly by online retailers. An increasing number of enterprise search vendors provide software for implementing faceted search applications.

My colleagues and I at Endeca are proud to be pioneers and innovators in this space. When we started in 1999, facets were an information science term familiar to only a niche of library scientists. Today, it is safe to say that, at least for site search and within the enterprise, faceted search has rapidly grown from a novelty to an expectation. Perhaps web search isn’t that far behind.

Interested in learning more? Check out the faceted search entry on Wikipedia, or read my blog, The Noisy Channel, which focuses more broadly on the challenges of enabling people to interact with information.

Daniel Tunkelang
Chief Scientist
T  646.792.3506
M 917.715.1779
E  dt@endeca.com
W http://thenoisychannel.com/

Endeca
120 W. 45th St, 5th Floor
New York, NY 10036
www.endeca.com

find / analyze / understand

This entry was posted on Monday, October 27th, 2008 at 6:05 am and is filed under Guest Authors, Innovations. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.